mlx-whisper is the fastest way to do automatic speech recognition on a Mac with OpenAI's Whisper models. In this video, we'll learn how to use it to transcribe an episode of the Tech Meme Ride Home podcast. We'll then see how well an LLM can answer questions about the episode based on the transcript and we'll see how fast mlx-whisper is compared to insanely-fast-whisper.
#openai #asr #mlxwhisper
00:00 insanely-fast-whisper isn't the fastest
00:16 Intro to mlx-whisper
00:30 mlx-whisper options and overview
01:05 Transcribing with mlx-whisper
01:41 Text version of transcript
01:58 JSON version of transcript
02:21 srt version of transcript
02:43 Summarising the transcript with llama3.2/Ollama
03:28 mlx-whisper vs insanely-fast-whisper
🟡 Code - https://github.com/mneedham/LearnDataWithMark/tree/main/mlx-whisper-playground
🟡 mlx-whisper - https://pypi.org/project/mlx-whisper/
🟡 mlx-whisper models - https://huggingface.co/collections/mlx-community/whisper-663256f9964fbb1177db93dc
🟡 llm - https://pypi.org/project/llm/